Files
clang-p2996/llvm/test/CodeGen/AMDGPU/atomic_optimizations_buffer.ll
Sameer Sahasrabuddhe 7a101798b7 Revert "[AMDGPU] Mark mbcnt as convergent"
This reverts commit 37114036aa.

The output of mbcnt does not depend on other active lanes, and hence it is not
convergent. The original change was made as a possible fix for

https://github.com/ROCm-Developer-Tools/HIP/issues/3172

But changing mbcnt does not fix that issue.

Reviewed By: ruiling, foad, yaxunl

Differential Revision: https://reviews.llvm.org/D153953
2023-06-30 13:10:44 +05:30

78 KiB